Sunday, October 2nd 2011

Sandy Bridge-E VT-d Broken In C1 Stepping, Fixed In C2 Stepping, Shortly After Launch

PC enthusiast customers and companies running corporate datacentres looking to buy into the new Sandy Bridge-E platform may want to wait a little while before handing over their hard-earned money to Intel. The initial batch of C1 revision Sandy Bridge-E processors have a bug - "errata" in Intel terminology - in them with VT-d, which means that hardware accelerated virtualization doesn't work properly with them (software only mode is unaffected). The feature when working properly, allows all hardware acceleration to work on the hosted operating system (virtual machine). This would allow things such as hard drive controllers to work, plus applications such as high-powered 3D games, typically First Person Shooters, to run at nearly full speed and the full Windows Aero desktop to be displayed on the hosted OS, as the hardware features of the graphics card can be used. Therefore, working VT-d is a critical feature for these kinds of applications.
Production of the C1 stepping should have already started, or is about to start. However, the C2 stepping isn't expected until next year, as the qualification process isn't expected to be complete until the end of this year.

Intel will only certify the Waimea Bay platform for PCI Express 2.0 at launch, since there aren't enough third party cards to test with. Some PCI Express 3.0 devices are still likely to work, but Intel doesn't guarantee compatibility. This doesn't bode well for the current 6-series motherboard with gen 3 switches actually working properly once cards and especially, Ivy Bridge processors arrive. Waiting for full qualification before purchasing is therefore advised, or an expensive motherboard replacement may be required in the not too distant future.

There is some good news however. For those wanting to continue using Windows XP, the Waimea Bay platform will be compatible with both 32-bit and 64-bit versions of the operating system. However, Intel won't be providing Rapid Storage drivers for 32-bit XP, which means that the basic Windows drivers will be handling drive access, which may not be optimal. Unsurprisingly, the upcoming Waimea Bay platform will also be supporting the upcoming Windows 8, when released.

It's not so good on the chipset front, either. Intel has increased the price of the X79 chipset by 20% compared to the X58 chipset, putting the X79 chipset somewhere in the region of $70. This is the list price, so the discounted or "street" price should be a lot lower. However, we are looking at a chipset that offers no new features over X58 and it's the same size as the 6-series chipsets - it should be cheaper to manufacture than the X58 chipset which only consists of the I/O Hub and the ICH10R. We wish AMD well with their new processors and chipsets, as competition can only be a good thing here and is clearly needed.

It's one thing for "errata" to be discovered some time after a product has been released and then to manufacture a revised processor, but it doesn't seem right for Intel to release processors with a known major feature fault like this, especially as most buyers are unlikely to know about it and Intel is even less likely to shout about it. Therefore, buyers would be wise to wait for the bug-fixed C2 version before upgrading, unless they are absolutely sure they won't need this virtualization feature. After all, are Intel going to offer a free replacement to the fixed version for customers of the initial C1 versions? Only this gesture can make releasing such an obviously flawed product right. Intel have offered such a free replacement recently with the faulty SATA controller, so there's hope yet they'll do the right thing. And just as importantly, it's worth keeping a close watch on what significant errata C2 might harbour, before upgrading.
Source: vr-zone.com
Add your own comment

49 Comments on Sandy Bridge-E VT-d Broken In C1 Stepping, Fixed In C2 Stepping, Shortly After Launch

#1
TRWOV
*sniff* *sniff* Barcelona? ;)
Posted on Reply
#2
the54thvoid
Intoxicated Moderator
I'll sit on my x58 until ivybridge i think. Just spent >£200 on W/C kit to play with my old i7 920. No upgrades for me this year. :)
Posted on Reply
#3
Ferrum Master
Yeah... mine 920 is also more than enough...

But seriously... this is a huge problem... it renders them pointless for server market...

Till X79 mates? :D
Posted on Reply
#4
newtekie1
Semi-Retired Folder
the54thvoidI'll sit on my x58 until ivybridge i think. Just spent >£200 on W/C kit to play with my old i7 920. No upgrades for me this year. :)
I'm still happy with my 4GHz Lynnfield.:laugh:

At least Intel caught it and let people know before the processors were released.
Posted on Reply
#6
qwerty_lesh
written with a strong degree if bias :banghead:

The express chipset may not contain a vast amount of feature differences over Tylersburg, but the platform on the whole is significantly better, sharing such praise wouldnt be beneficial to your hating tho wouldn't it? :slap:
Posted on Reply
#7
qubit
Overclocked quantum bit
qwerty_leshwritten with a strong degree if bias :banghead:

The express chipset may not contain a vast amount of feature differences over Tylersburg, but the platform on the whole is significantly better, sharing such praise wouldnt be beneficial to your hating tho wouldn't it? :slap:
There's been no bias and no hate.

If you want to buy a new system with these significant bugs in it, go right ahead, Intel is preparing to sell it to you. Everyone else will have been glad to have been informed of such a problem so that they can avoid getting stung by it. :rolleyes:
Posted on Reply
#8
crazyeyesreaper
Not a Moderator
correct ive actually starting using virtual machines and the like to test out some what ifs and potential problems also good for getting some older apps to work , so the fact Sandybridge E C1 silicon has a bug that makes that worthless im glad to know i didnt plan to use SB E but its still good to know about said problem.

but i would guess qwerty you love Intel so much any ill will or article must be biased right?

its a tech problem it was reported thats how this works, its also mentioned a fix is in place it just wont be ready for a short while. so dont get your panties in a bunch.
Posted on Reply
#9
D4S4
it would be nice if they sold them at lower prices because of this.
Posted on Reply
#10
seronx
NETBURST!! BARCELONA!!! AWFUL CPU!!!!

:p

Oh, well
Posted on Reply
#11
v12dock
Block Caption of Rainey Street
ESXi, Xen, Hyper-V all extremely popular hypervisors lets hope they get this issue fixed as soon as possible
Posted on Reply
#12
bear jesus
Damn :( part of why i am so interested in something with 6c/12t or 8c is the virtualization possibility's, i do agree though, it is very nice this info has come out before launch but i can't help but be unhappy that if sandy bridge E is my choice i would not get an upgrade this year.
Posted on Reply
#13
NC37
Just goes to show that just like the original Phenom was bugged, Intel is just as open to bugs as AMD.
Posted on Reply
#14
devguy
I'd have to guess that Intel will throw the majority of these C1 chips to OEMs, and tell them to not have an option to enable hardware VT-d in their BIOS/UEFI.

While I do agree that it is nice Intel is informing people before launch about this issue, it isn't so much like the "Barcelona" debacle you all reference. In that case, AMD helped motherboard manufacturers issue out a bios to fix the problem (albeit with about a 20% performance hit), and it really didn't affect any home consumer users. Intel motherboards cannot receive a BIOS update to fix this (most likely), nor the SATA issue on early 6 series chipsets.
Posted on Reply
#15
LAN_deRf_HA
This platform is looking more pointless everyday. It has no IPC benefit over 1155 despite twice the memory bandwidth. No pcie 3.0, and now no virtualization. It has 2 extra cores. That's it. Big whoop. You can fit two extra cores on 1155 with 22nm just fine. I get why they didn't learn their lesson with 1156 and 1366. 1366 came out first, people got attached to it and what not. This time though it's reversed. 1155 has a massive market saturation advantage, and so far 2011 isn't offering squat over it's cheaper already available cousin. If 2011 sells like shit just maybe Intel will go back to a singular consumer platform.

Unfortunately because 2011 is planned to have a very long life even if it didn't sell a single unit in the first year there's plenty of time for it to do well enough for Intel to consider it a success. Hell even if it was total flop I'm wondering if they still wouldn't try this divided platform crap again and again.
Posted on Reply
#16
Inceptor
LAN_deRf_HAUnfortunately because 2011 is planned to have a very long life even if it didn't sell a single unit in the first year there's plenty of time for it to do well enough for Intel to consider it a success. Hell even if it was total flop I'm wondering if they still wouldn't try this divided platform crap again and again.
They'll do it again and again. The 'enthusiast' high end socket has other customers, namely corporate and government customers looking for low-end workstations. It's useful to pull that aggregated segment out of the 'normal' low-end and mid-end segments. At least useful enough for a large corporation like Intel to make it financially viable, in some way.
Posted on Reply
#17
SteelSix
Well that's crappy news. I was waiting for X79. Plan B involves Z68 and 2500K; looks like that may have to do for a couple months.
Posted on Reply
#18
qubit
Overclocked quantum bit
SteelSixWell that's crappy news. I was waiting for X79. Plan B involves Z68 and 2500K; looks like that may have to do for a couple months.
Yeah, it's not good news, is it? The release of a proper, bug-fixed CPU is now so far away, that if someone wants to build a system now, they might as well just go ahead. Possibly waiting a short while to see what Bulldozer is like is the only thing worth waiting for. That's being released soon, but I can't remember the date, now.
Posted on Reply
#19
FordGT90Concept
"I go fast!1!11!1!"
I think there's a fair bit of exaggeration going on. Intel would not release a server-class processor without VT-d functionality (equivilent of shooting one self in the foot). It has a bug, but I highly doubt it is completely dysfunctional.
Posted on Reply
#20
qubit
Overclocked quantum bit
FordGT90ConceptI think there's a fair bit of exaggeration going on. Intel would not release a server-class processor without VT-d functionality (equivilent of shooting one self in the foot). It has a bug, but I highly doubt it is completely dysfunctional.
Yes, I've been wondering exactly how broken this is too. IMHO it's pretty bad for them to respin the chip.
Posted on Reply
#21
claylomax
qubitYeah, it's not good news, is it? The release of a proper, bug-fixed CPU is now so far away, that if someone wants to build a system now, they might as well just go ahead. Possibly waiting a short while to see what Bulldozer is like is the only thing worth waiting for. That's being released soon, but I can't remember the date, now.
Remember that AMD said Bulldozer would be released 90 days aproximately after the motherboards, so it's just around the corner.
Posted on Reply
#22
the54thvoid
Intoxicated Moderator
I think for anyone thinking of x79 for gaming rigs should really just wave it by. A 2600k and a top end gfx card will reap virtually the same benefits. So sure, you can go uber multi gfx but whether the cards run at x16 or x8 doesn't make a lot of real world diff (www.tomshardware.co.uk/pci-express-scaling-p67-chipset-gaming-performance,review-32164-9.html).

With Ivybridge coming in Spring (and AMD's reputed tweaks for BD) there will be more 'polished' tech available. PCI-e 3 will be more mature with perhaps some gfx vendors using it and you'll get tri-gate tech on IB.

Really, Intel screwed up by making SB so good. I bought X58 and SB cpu is better. You buy x79 now and in 6 months, IB will probably be better.

Nah, x79 seems like a waste to anyone other than those that really need it's functionality yet from all the web reports so far, it seems very hamstrung.
Posted on Reply
#23
DrPepper
The Doctor is in the house
Only problem is "hardware accelerated" VM's. The cpu's are fine for everyday stuff, no idea why some people think the thing might not work at all.
Posted on Reply
#24
Disparia
Well, that makes the decision easier. It was one thing to lose SAS support - future motherboards would have it and I could just swap it out. But with the importance of VT-d (VM's are my everyday stuff ;)), no reason to get SB-E now. I'll just have to wait for C2 along with SAS-equiped motherboards.

Hopefully they hold back SB-E Xeons. It is vitally important there, or the bug is overblown as Ford theorizes.


Come on AMD! I have nothing until 2012!
Posted on Reply
#25
Sihastru
qubitThe initial batch of C1 revision Sandy Bridge-E processors have a bug – “errata” in Intel terminology – in them with VT-d, which means that hardware accelerated virtualization doesn’t work properly with them (software only mode is unaffected). The feature when working properly, allows all hardware acceleration to work on the hosted operating system (virtual machine). This would allow things such as hard drive controllers to work, plus applications such as high-powered 3D games, typically First Person Shooters, to run at nearly full speed and the full Windows Aero desktop to be displayed on the hosted OS, as the hardware features of the graphics card can be used. Therefore, working VT-d is a critical feature for these kinds of applications.

It's one thing for "errata" to be discovered some time after a product has been released and then to manufacture a revised processor, but it doesn’t seem right for Intel to release processors with a known major feature fault like this, especially as most buyers are unlikely to know about it and Intel is even less likely to shout about it. Therefore, buyers would be wise to wait for the bug-fixed C2 version before upgrading, unless they are absolutely sure they won’t need this virtualization feature. After all, are Intel going to offer a free replacement to the fixed version for customers of the initial C1 versions? Only this gesture can make releasing such an obviously flawed product right. Intel have offered such a free replacement recently with the faulty SATA controller, so there's hope yet they'll do the right thing. And just as importantly, it's worth keeping a close watch on what significant errata C2 might harbour, before upgrading.
This part of the "news" is your own addition and has a nice scent of bias. You can say that it's your own analysis of the news bit on VR, but there are a few flaws with it.

The most important thing to note is that VT-d is not present on desktop platforms. Not only the CPU has to support VT-d, but also key motherboard components (NB/SB). On desktop consummer products these "features" are disabled. They are also disabled on desktop consummer CPUs.

Enterprise customers are not "most buyers" and they do know exactly what features they need and what they don't need. Also they would not normally buy a consumer grade desktop computer. They will buy enterprise grade workstations and servers, which is the Xeon brand, with Xeon compatible motherboards that will support VT-d if they indeed need this feature.

Games, like "first person shooters", are not an really what servers are for. Your trying to plant an idea in the minds of gullable TPU readers (I really hope they buy Bulldozers, I don't want them on the Intel camp), that SB-E has some major flaw that will affect their gaming performance. In reality this doesn't concern anyone since VT-d is disabled at the hardware level on consumer grade desktops.

So why is a bug in a disabled feature important? It's not.

By the time we get Xeons in the channel, they will all be C2. There will be no need for any recalls and "free replacements".

Sorry to lash out like this, but I thought TPU was above this.
Posted on Reply
Add your own comment
Apr 18th, 2024 06:15 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts